Generating Adversarial Examples With Conditional Generative Adversarial Net

Abstract

Recently, deep neural networks have significant progress and successful application in various fields, but they are found vulnerable to attack instances, e.g., adversarial examples. State-of-art attack methods can generate attack images by adding small perturbation to the source image. These attack images can fool the classifier but have little impact to human. Therefore, such attack instances are difficult to generate by searching the feature space. How to design an effective and robust generating method has become a spotlight. Inspired by adversarial examples, we propose two novel generative models to produce adaptive attack instances directly, in which conditional generative adversarial network is adopted and distinctive strategy is designed for training. Compared with the common method, such as Fast Gradient Sign Method, our models can reduce the generating cost and improve robustness and has about one fifth running time for producing attack instance.

6 pages

References in corpus (7)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Conditional Generative Adversarial Nets
• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• Conditional Image Synthesis With Auxiliary Classifier GANs
• Transferability in Machine Learning: from Phenomena to Black-Box Attacks using Adversarial Samples
• The Space of Transferable Adversarial Examples
• Certifying Some Distributional Robustness with Principled Adversarial Training